The expanded ability to collect data at all scales-molecular, cellular, tissue, organism and population-has created unparalleled opportunities for biomedical discovery. These opportunities cross all areas of research from basic science to clinical care. In response to these tremendously exciting emerging challenges in data science, Stanford University announced the creation of a new Department of Biomedical Data Science (DBDS) to begin in fall of 2015. Fundamental to the DBDS is bringing together faculty in (1) informatics and computer science, and (2) biostatistics and mathematical modeling, who work closely with a broad range of (3) biomedical science collaborators to advance knowledge. The Stanford Biomedical Informatics (BMI) training program is focused on the creation of new methods for the organization, analysis and modeling of biomedical data and knowledge. The BMI program has been a small interdisciplinary program at Stanford for more than 33 years; nonetheless, it has produced many leaders in biomedical informatics and data science. The BMI program will now have its administrative home in the DBDS, and will become the epicenter for biomedical data science training at Stanford. We are able quickly to respond to the shortage of trained scientists in biomedical data science because of a flexible curriculum, an unusually fertile set of course offerings, and a plethora of research opportunities. In this proposal, we outline a plan to engage faculty broadly across the University to create scalable mechanisms for training the next generation of biomedical data scientists, and creating a pathway for data science within the BMI program that stresses statistical reasoning, machine learning and data mining of biomedical data.